Journal of Chromatography A, 918 (2001) 303–310
www.elsevier.com / locate / chroma
Automated solid-phase extraction for sample preparation followed
by high-performance liquid chromatography with diode array and
mass spectrometric detection for the analysis of resveratrol
derivatives in wine q
´
´ C.G. Barroso
C. Domınguez*,
D.A. Guillen,
´
´ , Facultad de Ciencias, Universidad de Cadiz
´ , Polıgono
´
´ San Pedro s /n,
Rıo
Departamento de Quımica
Analıtica
´ ), Spain
11510 Puerto Real ( Cadiz
Received 20 November 2000; received in revised form 28 February 2001; accepted 7 March 2001
Abstract
A method has been developed for the simultaneous determination of resveratrol in all its forms (free isomers and
glycosylates) in wines by high-performance liquid chromatography with diode array and mass spectrometric detection. Prior
to injection into the column, preconcentration of the sample by automated solid-phase extraction is carried out. In the
detection by UV absorption, quantitation was carried out at 280 and 305 nm, and in detection by mass spectrometry,
quantitation was performed in the selected ion monitoring mode at m /z 228 and at m /z 238. A comparative study between
both detection systems was carried out.  2001 Elsevier Science B.V. All rights reserved.
Keywords: Sample preparation; Wine; Food analysis; Resveratrol; Glycosides; Phenolic compounds
1. Introduction
Resveratrol (3,5,49-trihydroxystilbene) and its derivatives are natural phenolic compounds found in
many families of plants [1–8]. The grape and related
products, such as wine, are probably the food
products that contain these compounds in greater
proportions.
q
Presented at the 29th Scientific Meeting of the Spanish Group
of Chromatography and Related Techniques, Alcala´ de Henares
(Madrid), 12–14 July 2000.
*Corresponding author. Tel.: 134-956-016-363; fax: 134-956016-288.
´
E-mail address: [email protected] (D.A. Guillen).
The presence of these stilbenes in grapes and
derived products has been under study for some
years now, for two main reasons. First, it has been
demonstrated that these compounds act as phytoalexins in the plant [9–15], being synthesized in response
to situations of stress, such as attack by pathogens,
UV radiation, or lesion. Second, it has also been
observed that they are beneficial for human health
through their antioxidant properties, since it seems
they may be partially responsible for the lower
mortality rates from cardiovascular disease observed
in populations that customarily consume wine in
moderation [15–19]. Further, it has recently been
published that it is possible these compounds possess
certain chemico-preventive activity against cancer
[20–22], as a selective modulator of estrogen re-
0021-9673 / 01 / $ – see front matter  2001 Elsevier Science B.V. All rights reserved.
PII: S0021-9673( 01 )00748-8
304
´
et al. / J. Chromatogr. A 918 (2001) 303 – 310
C. Domınguez
ceptors [43], as well as having possible application
in pathologies of the nervous system [44].
Many different methods have been developed for
determining resveratrol in wines. The most common
are based on high-performance liquid chromatography (HPLC), using detection by UV absorption
[23–26], by electrochemical [27,28] or fluorometric
[29,30] methods, or by gas chromatography (GC)
[31,32] and GC–mass spectrometry (GC-MS) [33–
35] techniques. Recently the introduction of hyphenated techniques, such as LC–MS, have been shown
to be powerful tools for the resolution of complex
samples, as in this case. As a result, this technique
has already been applied for the determination of
phenolic compounds in foodstuffs [36].
In respect of the preparation of the sample for its
analysis, some authors do not do this, but rather
inject the sample directly [23,25,27–30], although
the majority use some kind of sample preconcentration and cleaning stage before its injection, such
as a liquid–liquid extraction [37–39] or a solid-phase
extraction (SPE) [26,40].
The reason for the success of SPE lies in the
numerous advantages it offers, such as a high
selectivity, fast speed and facility of automation of
the procedure [41,42].
This study presents the development of a method
for determining simultaneously all the monomers of
resveratrol in wine, by the application of HPLC with
diode array (DAD) and MS detection. Prior to the
injection in HPLC, a stage of preconcentration of the
sample by automated SPE is performed, employing a
polymeric adsorbent (polystyrene–divinylbenzene)
which gives better results than C 18 since it has a
better capacity for retention of compounds with a
wide range of polarities. The separation by elution
gradient in HPLC and the conditions of MS detection
were optimized with the aim of achieving wellresolved peaks and the maximum possible signal in
the mass detector.
2. Experimental
2.1. Reagents and standards
The acetonitrile, methanol and tetrahydrofuran, of
HPLC quality, and the other reagents were supplied
by Scharlau (Barcelona, Spain). The trans-resveratrol was acquired from Sigma (St. Louis, MO, USA)
and the 3,4,5-trimethoxycinnamic acid, employed as
internal standard, from Fluka (Buchs, Switzerland).
The water employed was previously purified in a
Milli-Q system (Millipore, Bedford, MA, USA). The
aqueous solutions were filtered using cellulose acetate membranes of 0.45 mm (47 mm) (Micron
Separation, Westboro, MA, USA) and were degasified in an ultrasonic bath before being used.
2.2. Samples
A variety of different types of wines were analyzed: red Rioja wines, port and fino sherry wine
´
supplied by Bodegas Osborne (El Puerto Sta. Marıa,
´
Cadiz,
Spain), and double-macerated reds and white
table wines, produced in Spain and acquired commercially.
2.3. Instrumentation and conditions
The analyses were performed using a Waters
Integrity HPLC–DAD–MS system (Waters, Millipore, Milford, MA, USA). The chromatographic
system consisted of a Model 616 pump and a Model
600S gradient controller, to which were connected a
Model 717 plus autosampler, a Model 996 detector
of aligned photodiodes and a Thermabeam MS
detector with modified particle beam interface,
equipped with a source of ionization by electron
impact (70 eV), a quadrupole analyzer and an
electron multiplier detector.
The separation was performed using a Symmetry
C 18 column (15032.1 mm I.D., 5 mm particle size).
The control of the equipment, and the acquisition
and treatment of data was performed with the
Millenium 2010, version 2.21, software.
The chromatographic conditions were the following: 0.2 ml / min flow-rate; 50 ml injection volume;
eluent A was distilled water (Milli-Q quality),
adjusted to pH 2.5 with sulfuric acid; and eluent B
was pure acetonitrile; the elution gradient is shown
in Table 1.
Detection by UV absorption was performed by
sweeping between 240 and 305 nm, with a resolution
of 1.2 nm, and the quantitation was done at 280 and
305 nm. The detection and quantitation by MS was
´
et al. / J. Chromatogr. A 918 (2001) 303 – 310
C. Domınguez
Table 1
Elution gradient programme
305
2.5. Calibration
t (min)
A (%)
B (%)
Curve
0
60
90
100
110
100
50
50
0
100
0
50
50
100
0
–
7
7
7
7
performed in the selected ion monitoring (SIM)
mode, at m /z 228 (for detection of isomers of
resveratrol) and at m /z 238 (for detection of the
internal standard). The conditions imposed in the
interface, previously optimized by means of simplex,
were the following: source temperature, 2208C;
nebulization temperature, 858C; expansion region
temperature, 758C, helium flow-rate, 250 ml / min.
2.4. Preparation of the samples
The samples were filtered through nylon filters of
0.45 mm (13 mm) (Osmonics, Minnetonka, MN,
USA) before subsequently undergoing an SPE, for
the purpose of cleaning and preconcentrating them
before injection into the HPLC–DAD–MS system.
The SPE stage was performed by a totally completely automated method using a semi-flexible and
automatic robotic system: a Benchmate work station
(Zymark, Hopkinton, MA, USA). The cartridges
used were LiChrolut EN (Merck, Darmstadt, Germany), containing 200 mg of the polymeric adsorbent, polystyrene–divinylbenzene. These were
conditioned first with 5 ml of methanol and followed
by 3 ml of water. A sample of 5 ml of wine to which
123 ml of 216 mg / l solution of 3,4,5-trimethoxycinnamic acid as internal standard has been added
was diluted 1:1 with water and 9.8 ml of diluted
sample was loaded in the cartridge previously conditioned. Then the column was rinsed with 0.6 ml of
water and later dried with helium for 150 s. Finally
the compounds of interest were eluted with 1.2 ml of
tetrahydrofuran and followed by 1.2 ml of water.
The flow-rates applied during the automated procedure are the following: condition flow, 0.25 ml / s;
load flow, 0.01 ml / s; wash flow, 0.05 ml / s; elution
flow, 0.05 ml / s; air flow, 0.10 ml / s; air factor, 0.6.
All the steps and flow-rates of the automated SPE
procedure were previously optimized by the authors.
For the trans-resveratrol, the calibration curves
were drawn for the detection both by UV absorption
and by MS, from various standard solutions prepared
by diluting a stock solution of 100 mg / l of transresveratrol in aqueous methanol at 60%, such that
they cover a range of concentration between 2.5 and
50 mg / l. All the solutions were stored at 48C and
protected from light.
For the calibration of the cis-resveratrol, since its
commercial standard is not available, a fraction of
the stock solution of 100 mg / l trans-resveratrol
prepared for its own calibration was taken and then
irradiated for 15 min in a climatic chamber with
control of temperature, humidity and illumination. In
this chamber, by a method of accelerated environmental ageing using a solar radiation panel by xenon
(1500 W) and at 288C and 92% humidity, the transresveratrol is converted into cis-resveratrol.
The lower level observed in the peak corresponding to trans-resveratrol in the irradiated solution was proportional to the height of the new peak
which appeared following the trans-resveratrol, corresponding to cis-resveratrol, which was identified
by its UV absorption and mass spectra. Since no
other peaks were detected under these conditions, the
concentrations of cis-resveratrol were assigned on
the basis of the reduction observed for trans-resveratrol after its irradiation.
All the samples and calibration standards were
quantified by the internal standard method. The
compound used as internal standard, 3,4,5-trimethoxycinnamic acid, was selected from the collection of
polyphenolic compounds maintained by the research
group, bearing in mind that this compound was not
present in the samples of wine under study. This
compound was added at a fixed and known concentration to all the samples from the commencement of the analysis procedure.
3. Results and discussion
3.1. Chromatographic procedures
First the chromatographic conditions were optimized to ensure that the compounds of interest were
306
´
et al. / J. Chromatogr. A 918 (2001) 303 – 310
C. Domınguez
well resolved. For this, the various elution phases
and corresponding chromatographic gradient were
studied in such a way that, as well as achieving an
appropriate resolution, a better signal was obtained
in the mass detector; this is because the composition
of the phases in which the compounds of interest are
eluted considerably affects the yield obtained in the
transference of the analyte from the liquid phase to
the mass spectrometer. The optimum conditions
reached are those described in the preceding section.
Next the stage of preconcentration by automated
SPE was studied. Taking as the starting point the
conditions previously devised by the authors for the
determination of the phenolic compound content of
sherry wine [40], the conditions and the instrumental
parameters were slightly modified to adapt the
method to other types of wine. Fig. 1 shows the
chromatograms obtained with UV detection at 305
nm from the direct injection of a red wine (Fig. 1a)
and following the stage of automated SPE (Fig. 1b);
as can be observed, the increase of the signals
Fig. 1. Chromatograms at 305 nm, corresponding to a red wine.
(a) By direct injection. (b) After the stage of automated SPE.
Peaks: 15trans-resveratrol glycoside, 25cis-resveratrol glycoside, 35trans-resveratrol, 45cis-resveratrol.
corresponding to the compounds of interest is significant. This result is important particularly bearing in
mind that the system of MS detection used in this
present study (LC–MS interface; particle beam;
ionization by electron impact) has a fairly limited
sensitivity. These assays were performed with real
samples, with their peaks being identified by both
their UV spectra absorption and their mass spectra.
3.2. Calibration
Considering that only the trans isomer of resveratrol is commercially available as a standard, the
possibility of obtaining a calibration curve for each
of the species under study is rather limited. There are
two alternative solutions to this difficulty described
in the bibliography: one is the calculation of the
concentrations of the different compounds using
exclusively the calibration curve of trans-resveratrol;
the other is the generation of the cis isomer by
means of irradiation of a standard solution of transresveratrol and then quantifying the amount of cisresveratrol thus formed in function of the quantity of
trans-resveratrol seen to have ‘‘disappeared’’.
The first of these options is arguably inadequate,
especially considering the spectral characteristics of
the two compounds. Fig. 2 shows the UV absorption
and mass spectra for the two isomers of resveratrol.
The mass spectra is the same for both isomers, but
the UV absorption spectra are clearly fairly different:
the trans isomer presents an absorption maximum at
305.7 nm (e 526 000) while that presented by the cis
isomer is 287.8 nm (e 512 000); consequently there
would be a considerable error in quantifying cisresveratrol by means of the curve of the trans
isomer.
This is not the case with the glycosides since the
spectra of the combined forms are similar to those
corresponding to the free forms. Consequently, it
would be reasonable, in order to perform a quantitation with the least possibility of error, to have
available two calibration curves: one to quantify
trans-resveratrol and its glycoside, the other for the
cis isomer and its glycoside.
This is the second alternative described in the
bibliography, in which the cis isomer is generated by
solar light irradiation, from a solution of transresveratrol of known concentration; the degree of
´
et al. / J. Chromatogr. A 918 (2001) 303 – 310
C. Domınguez
Fig. 2. (a) UV absorption spectra (normalized at absorbance
maxima) of trans-resveratrol. (b) UV absorption spectra (normalized at absorbance maxima) of cis-resveratrol. (c) Mass spectra
(normalized at maximum intensity) of both isomers of resveratrol.
307
conversion is then quantified and the corresponding
calibration curve for this isomer is obtained by
dilution of the initial solution. However, the procedure followed to achieve these standards by irradiation involves certain errors, which could be
avoided if cis-resveratrol were available commercially as a standard.
The calibration of cis-resveratrol and its glycoside
obtained from the irradiated standards of trans-resveratrol provides good analytical characteristics, as
can be observed from Table 2. However, taking into
account the similarity between the mass spectra of
the two isomers of resveratrol and their corresponding glycosides, the possibility is still open of
using a single calibration curve, in the case of
performing the quantification by MS. But for this
alternative to be considered valid, the results for the
concentration of cis-resveratrol calculated using the
curve of trans-resveratrol should not differ significantly from those obtained by applying the curve for
cis-resveratrol constructed with the irradiated standards of trans-resveratrol. To assess whether or not
there is such a difference, the t-test for means of two
paired samples, for 95% confidence, was applied to
an identical series of samples, making a comparison
of the concentrations of cis-resveratrol calculated
from its own curve and those calculated from the
curve of trans-resveratrol. The result was that there
are no significant differences between the results
obtained.
Thus by constructing a single calibration curve,
from the synthetic compound, it is possible to
quantify the four monomers when the detection
system used is mass spectrometry. The analytical
parameters, obtained from the calibration curves of
Table 2
Analytical parameters of the method
trans-Resveratrol
Concentration range(mg / l)
r2
Standard error
Detection limit
Quantification limit
Analytical sensitivity
cis-Resveratrol
MS (228 m /z)
DAD 305 nm
MS (228 m /z)
DAD 280 nm
2.5–50
0.9980
0.2324
1.202
4.005
0.4625
2.5–50
0.9997
0.3934
0.948
3.16
0.3533
1.23–40.06
0.9972
1.2434
2.558
8.525
0.9844
1.23–40.06
0.9997
0.2614
0.834
2.781
0.3211
´
et al. / J. Chromatogr. A 918 (2001) 303 – 310
C. Domınguez
308
trans- and cis-resveratrol by means of the two
detection systems used, is given in the Table 2. It can
be seen that both detection systems give similar
results.
3.3. Recovery and repeatability
The recovery and repeatability of this analytical
method has been studied, by means of the processing
of samples of the same wine six times; a known
quantity of trans-resveratrol had previously been
added to this wine. In this way, a series of six results
were obtained for each detection system; these are
shown in Table 3.
It can be observed that the method provides an
acceptable relative standard deviation (RSD) for the
four compounds with both systems of detection,
since the dispersion of the data is not very wide.
Although the highest value of RSD is obtained for
trans-resveratrol with MS detection, the mean value
of concentration obtained by this system is closer to
the known figure than that resulting from the detection by UV absorption, in spite of quantifying it at
the wavelength of maximum absorption of this
compound. It is also observed that for the spiked
species, the recovery of the method of analysis is
closer to reality when measured by the MS detection
system than when measured by the UV absorption
system. This could be an indication that for the latter
system there exists a systematic error in the method.
In contrast, the quantification by UV absorption is
more accurate than by MS.
In addition it is notable that in the MS detection
system, the lower RSD corresponds to cis-resveratrol, the compound present in the lowest concentration in the sample studied. This is explained by
the high selectivity of this system of detection, which
Fig. 3. Chromatograms of an extract of red wine. (a) Detection by
UV absorption at 305 nm, (b) detection by MS in the SIM mode,
at m /z 228.
provides simpler chromatograms that are easier to
integrate since no other compounds appear co-eluted
with those of interest, thus avoiding a quantitation
error.
In Fig. 3, the chromatograms obtained by UV
detection (Fig. 3a) and by MS (Fig. 3b) of the same
sample are presented. The simplicity and cleanness
of the chromatogram provided by MS, in comparison
Table 3
Repeatability study of a red wine spiked with 15 mg / l of trans-resveratrol (six repetitions)
Glycoside
Mean (mg / l)
SD
RSD (%)
Recovery (%)
Glycoside
trans-Resveratrol
cis-Resveratrol
trans-Resveratrol
cis-Resveratrol
MS
DAD
MS
DAD
MS
DAD
MS
DAD
1.302
0.141
10.846
–
1.598
0.044
2.753
–
1.945
0.160
8.219
–
1.909
0.047
2.456
–
15.606
2.024
12.970
98.83
11.278
0.744
6.594
71.42
0.838
0.034
4.018
–
0.720
0.063
8.775
–
´
et al. / J. Chromatogr. A 918 (2001) 303 – 310
C. Domınguez
with that obtained by UV absorption at 305 nm, is
very evident. This difference in selectivity between
the two systems of detection is due to the fact that, in
a sample of wine, there is a greater chance of finding
other compounds, apart from those of interest, which
absorb at 305 nm than other compounds with
molecular fragments of m /z 228.
309
than 1 would indicate the existence of a systematic
error in one of the methods.
Lastly, the method of analysis was applied to the
four derivatives of resveratrol in various types of
wine, and the results obtained are shown in Table 4.
4. Conclusions
3.4. Comparative study of the two systems of
detection
To compare the two systems of detection and to
check whether one of them shows any systematic
error, the values obtained by both systems for a
series of samples composed of different types of
wine fortified with trans-resveratrol were analyzed
by means of regression. A regression line was
constructed in which on the x-axis were represented
the results of detection by UV absorption and on the
y-axis the results of detection by MS; an r value of
0.9912 was obtained for this line. The slope (b) and
intercept (a), together with the confidence limits for
95%,
were:
b51.735560.042
and
a52
2.004862.301, which indicates, with respect to the
ideal values (b51, a51), that no significant differences exist in the case of the intercept, whereas they
do exist in the case of the slope, since a value greater
The development of hybrid instrumental techniques widens the alternatives for analyzing compounds present in complex samples and at low
concentrations. This is the case of the analysis of
resveratrol in wines.
The method of analysis of resveratrol in wines
proposed in this study offers many advantages. First,
the SPE stage provides good repeatability and selectivity, speed, minimum manipulation of samples and,
furthermore, is a procedure that can be fully automated. Second, the quantitation of the four monomers of resveratrol by means of mass spectrometry,
employing a single calibration curve constructed
from internal standards prepared from a commercially-available standard, simplifies the procedure
and offers more reliability. Third, detection by mass
spectrometry, being more selective, permits simpler,
clean chromatograms to be obtained which facilitate
the quantitation.
Table 4
Concentrations (mg / l) found in different types of wine
Red wine – 1
Red wine – 3
Red wine – 4
Red wine – 5
Red wine – 6
Red wine – 9
Red wine – 10
Red port wine
Double macerated red wine – 1
Double macerated red wine – 2
Double macerated red wine – 3
White table wine
Sherry wine
Pedro Ximenez wine
Muscatel wine
trans-Resveratrol
glycoside
cis-Resveratrol
glycoside
trans-Resveratrol
cis-Resveratrol
MS
DAD
MS
DAD
MS
DAD
MS
DAD
0.965
0.727
0.840
1.543
1.223
1.300
0.727
1.281
6.968
7.354
4.078
n.d.
n.d.
n.d.
n.d.
1.893
0.418
1.447
2.300
1.882
1.912
0.418
1.812
9.442
8.504
5.076
n.d.
n.d.
n.d.
n.d.
1.191
0.920
0.918
1.040
1.194
1.315
0.765
2.484
7.816
10.553
6.944
n.d.
n.d.
n.d.
n.d.
1.656
1.524
0.909
1.548
1.505
1.626
1.015
2.222
5.301
7.091
5.002
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
1.073
0.836
n.d.
2.009
2.000
1.387
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
1.274
0.787
n.d.
2.157
1.844
1.439
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
1.174
1.295
n.d.
2.758
2.159
1.917
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
1.024
1.065
n.d.
1.660
1.482
1.315
n.d.
n.d.
n.d.
n.d.
310
´
et al. / J. Chromatogr. A 918 (2001) 303 – 310
C. Domınguez
Acknowledgements
This study was supported by the CICYT of the
Spanish Ministry of Education and Science and
FEDER of U.E. (project 1FD97-0683.01). The authors are indebted to Bodegas Osborne & Cia. for
supplying samples.
References
[1] J.L. Ingham, Phytochemistry 15 (1976) 1791.
[2] R.G. Powell, M.R. Tepaske, R.D. Plattner, J.F. White, S.L.
Clement, Phytochemistry 35 (1994) 335.
[3] R.J. Kumar, D. Jystostna, G.L.D. Krupadanam, G. Srimannarayana, Phytochemistry 27 (1988) 3625.
[4] J. Gorham, S.J. Coughlan, Phytochemistry 19 (1980) 2059.
[5] V.S. Sobolev, R.J. Cole, J.W. Dorner, J. AOAC Int. 78
(1995) 1177.
¨
¨
[6] R. Hain, B. Bieseler, H. Kindl, G. Schroder,
R. Stocker,
Plant Mol. Biol. 15 (1990) 325.
[7] P. Langcake, R.J. Pryce, Phytochemistry 16 (1977) 1193.
[8] P. Langcake, A. Cornford, R.J. Pryce, Phytochemistry 18
(1979) 1025.
[9] R. Bessis, P. Jeandet, M. Adrian, A.C. Breuil, S. Debord,
Rev. Oenologues 85 (1997) 5.
[10] R.J. Pryce, P. Langcake, Phytochemistry 16 (1977) 1452.
[11] M. Barlass, R.M. Miller, T.J. Douglas, Am. J. Enol. Vitic. 38
(1987) 65.
˜
˜ A.R.
´ J.M. Zapata, R. Munoz,
[12] A.A. Calderon,
M.A. Pedreno,
´ News Phytol. 124 (1993) 455.
Barcelo,
´
[13] J.P. Roggero, M.C. Garcıa-Parrilla,
Sci. Aliments 15 (1995)
411.
[14] J.P. Blond, M.P. Denis, J. Bezard, Sci. Aliments 15 (1995)
347.
[15] F. Bravo, Alimentaria, En-Febr (1996) 71.
´ A. Gonzalo, J.M. Ramon,
´ S. Mınguez,
´
[16] I. Hurtado, P. Caldu,
C. Fiol, J. Agric. Food Chem. 45 (1997) 1283.
[17] C.S. Stockley, in: B. Rankine (Ed.), Information Report,
Austr. Soc. Wine Educ. (1995) 1.
¨
[18] P. Furst,
Proc. Nutr. Soc. 55 (1996) 945.
[19] E.N. Frankel, A.L. Waterhouse, P.L. Teissedre, J. Agric.
Food Chem. 43 (1995) 890.
´ 30 (1995) 224.
[20] S. Brun, Cah. Nutr. Diet.
[21] M. Jang, L. Cai, G.O. Udeani, K.V. Slowing, C.F. Thomas,
C.W.W. Beecher, H.H.S. Fong, N.R. Farnsworth, A.D. Kinghorn, R.G. Mehta, R.C. Moon, J.M. Pezzuto, Science 275
(1997) 218.
[22] M.C. Cravero, V. Dell’Oro, Sevi No. 2557 (1995) 2759.
´ A.I. Romero-Perez,
´
[23] R. Lamuela-Raventos,
A.L. Waterhouse,
M.C. de la Torre-Boronat, J. Agric. Food Chem. 43 (1995)
281.
[24] M. Fregoni, L. Bavaresco, D. Petegolli, M. Trevisan, C.
Ghebbioni, Vignevini 6 (1994) 33.
[25] D.M. Goldberg, E. Ng, A. Karumanchiri, J. Yan, E.P.
Diamandis, G.J. Soleas, J. Chromatogr. A 708 (1995) 89.
´ Equipos Tecnologıa
´
[26] A. Gonzalo, P. Vidal, Alimentacion,
October (1995) 67.
[27] K.D. McMurtrey, J. Minn, K. Pobanz, T.P. Schultz, J. Agric.
Food Chem. 42 (1994) 2077.
[28] K.D. McMurtrey, in: T.R. Watkins (Ed.), Wine Nutritional
and Therapeutic Benefits, ACS Symposium Series 661
(1997) 45–55.
[29] R. Pezet, V. Pont, P. Cuenat, J. Chromatogr. A 663 (1994)
191.
[30] T. Okuda, K. Yokotsuka, Am. J. Enol. Vitic. 47 (1996) 93.
[31] G. Revel, T. Hogg, C. Santos, J. Int. Sci. Vigne Vin 30
(1996) 31.
[32] A. Antonelli, C. Fabri, G. Lercker, Chromatographia 42
(1996) 469.
[33] D.M. Goldberg, J. Yan, E. Ng, E.P. Diamandis, A. Karumanchiri, G.J. Soleas, A.L. Waterhouse, Anal. Chem. 66 (1994)
3959.
[34] G.J. Soleas, D.M. Goldberg, E.P. Diamandis, A. Karumanchiri, J. Yan, E. Ng, Am. J. Enol. Vitic. 46 (1995) 346.
[35] D.M. Goldberg, J. Yan, E. Ng, E.P. Diamandis, A. Karumanchiri, G.J. Soleas, A.L. Waterhouse, Am. J. Enol. Vitic. 46
(1995) 159.
[36] A. Gioacchini, A. Roda, G.C. Galleti, J. Chromatogr. A 730
(1996) 31.
[37] E. Celotti, R. Ferrarini, R. Zironi, S.L. Conte, J. Chromatogr.
A 730 (1996) 47.
[38] V. Vacca, G. Madau, M.A. Franco, P. Fenu, Riv. Sci.
Alimentazione 4 (1995) 565.
[39] F. Mattivi, Z. Lebensm. Unters.-Forsch. 196 (1993) 522.
´ C.G. Barroso, J.A. Perez-Bus´
[40] C. Chilla, D.A. Guillen,
tamante, J. Chromatogr. A 750 (1996) 209.
´
´
[41] D.A. Guillen,
C.G. Barroso, J.A. Perez-Bustamante,
J.
Chromatogr. A 730 (1996) 39.
´ Doctoral Thesis, University of Cadiz, Cadiz,
[42] D.A. Guillen,
1994.
[43] J. Pezzuto, K.P.L. Bhat, Sevi No. 2782 / 83 (1999) 4169.
[44] A. Bertelli, Sevi No. 2782 / 83 (1999) 4189.